Carl Pry, a Senior Advisor at Treliant, is a seasoned executive with banking law, corporate finance, and regulatory compliance experience in Fortune 500 institutions, regional banks and industry consulting firms. Carl advises clients on commercial compliance, fair lending, corporate treasury and risk management. Over the last 27 years, Carl has…
Few topics have generated as much attention this year than the new 2018 HMDA rules, from the expanded definition of a HMDA-reportable application (any dwelling-secured loan or line of credit is now covered) to the vast increase in data that must now be collected and submitted. In order to implement these changes, the effort cannot be limited to just compliance personnel. Operations, loan officers, IT, vendors, and others are critical to the effort to have a submission-ready file by next March 1st. Banks have spent countless hours trying to get things right (see how I worked that in there?) to satisfy not just regulators but because banks, and the industry as a whole, will be judged publicly by the results. There will be so many more data points upon which to parse a bank’s fair lending performance, increasing the possibility of fair lending criticism exponentially.
But on this journey of implementation, several critical areas of confusion and concern have been noted. How is your bank doing? Will you be ready to submit good data by March 1st? The following are a few of those problem areas, issues to be aware of, and other pain points to address.
How do you define a HMDA-reportable application?
The Bureau of Consumer Financial Protection (Bureau), as they call themselves now, declined to offer any additional clarity to the definition of an “application” in the final rule. They left the definition as is, meaning there continues to be flexibility in defining what is reportable. The definition covers requests for a HMDA-covered loan “made in accordance with procedures used by a financial institution for the type of credit requested.” That language is very close to that in Regulation B, but not identical. This means it is possible to have an “application” under Reg. B but not under HMDA. It also means the action taken on that “application” may differ; for instance, a situation where an applicant is fully approved but the property has issues would necessitate a denial under Reg. B, but the action taken under HMDA would be Approved Not Accepted.
The definition of an application under the TILA RESPA Integrated Disclosures (TRID) rule in Reg. Z is different still. The Bureau made it clear that the HMDA and TRID definitions are not the same, which again means you may have a HMDA-reportable application that doesn’t rise to the level of triggering the TRID requirements, or vice-versa. All this creates opportunities for mischief – how do you define an application under each of those rules? Whether they are the same or different, be prepared to articulate your position. Not only are you doing this for HMDA reportability reasons, but knowing when adverse action notices are necessary, when Loan Estimates are required, etc., is critical. Take this opportunity to clarify for your bank what an application is (as well as when it is submitted and/or received) under each rule so you don’t find yourself complying with HMDA at the expense of Reg. B or TRID.
New data element questions
Several of the new data elements beg for clarification from the Bureau. The best advice for now is to decide how you’ll report this information, then be consistent in doing so.
Manufactured home communities. Under the new rule, a loan must be secured by a dwelling to be reportable. This change means non-dwelling secured home improvement loans (which were reportable in 2017 and prior) are no longer reportable, but home equity loans having a non-improvement purpose are reported. However, a loan secured by a manufactured home community (meaning, for instance, a mobile home park where the collateral is just the lots) is considered secured by a dwelling, and is therefore HMDA-reportable, even though there are no physical dwellings taken as collateral. But then consider what the proper construction method (another data point) would be for these loans. Verbal guidance from the Bureau states that this would be “manufactured,” but that’s not as clear as it could be. Logic suggests this should be “NA,” but for now the recommendation is “manufactured.”
Employee applications. Now as before, for privacy reasons, income is not reported if the applicant is an employee. However, CLTV, DTI, and credit score are not similarly exempted. One might think the same privacy logic might apply for these fields, but for now they are reported even if the applicant is an employee.
Automated Underwriting Systems. Banks must report the results from automated underwriting systems (AUS), but sometimes results from multiple AUSs are obtained. Which result must be reported if this is the case? It’s easy to say the one that primarily drove the credit decision, but further guidance is needed here as well. But for now, if multiple AUSs are used, identify the result that most drove your credit decision and report that one.
If the bank provides a counteroffer to an applicant and he/she says no thanks, it is reported as a denial. This has not changed from earlier guidance. If the applicant says OK to the counteroffer but the loan does not close, the loan amount reported is the amount counteroffered. However, all the action taken codes are still possibilities – this is the change from previous guidance. If the applicant “accepts” the counteroffer but the loan does not close, it could be considered Approved Not Accepted, but this is not always the case. If the applicant was fully approved (meaning it was not a conditional approval) and the loan does not close it would be Approved Not Accepted, but if the counteroffer was conditional, Approved Not Accepted would not be the correct choice. Close attention must be paid to counteroffer situations; exactly what occurs must be closely monitored so that the correct Action Taken code can be recorded.
What if my bank is impacted by the Dodd-Frank reform law?
The Economic Growth, Regulatory Relief, and Consumer Protection Act (also known as S. 2155), signed into law in May of 2018, provides that banks that originate fewer than 500 open-end or 500 closed-end loans in each of the previous two calendar years are exempt from HMDA’s expanded data submission requirements for those loan types. At first glance this looks like great news, but do not misunderstand what this means. It does not mean that the bank is exempt from HMDA reporting entirely; it merely means that the new expanded data elements introduced for 2018 need not be reported. In other words, for 2018 the bank goes back to reporting the information it did for 2017 and previously.
But there are a few important issues to consider. First, many smaller banks that sell their originated loans into the secondary market are finding that investors are requiring the originating bank (whether exempt or not) to report all the data to it, including the new 2018 elements. The new rule allows investors to report some, but not all, data as “NA” (or similar coding) for purchased loans. Some investors are requiring banks to provide it all the data regardless of the exemption so it can monitor its own fair lending performance. Thus the small bank doesn’t get the relief it anticipated from the Dodd-Frank reform provisions.
Speaking of fair lending performance, might you consider continuing to collect the new data even if your bank is now exempt? After going through all the work of setting up your systems to collect the new information, why stop it now? It will have great utility in your fair lending analytics by allowing you a much deeper look into your bank’s decisioning and pricing practices. But this is more a control decision rather than a compliance one; each bank must decide for itself.
Fair lending impacts
Do not underestimate the impact so much additional data will have on fair lending analyses and analytics. This is not just a concern about what your regulators can do with your data; the opportunity to parse so much more information to identify possible disparities and hot spots is eagerly anticipated by the press and public interest groups alike. Take this time to understand what your data says about your bank’s fair lending profile; before you know it, the information will be available for public consumption.
Keep in mind another important change to HMDA, one that is not data-related: your bank’s 2018 and future LARs will (as your 2017 LAR is now) be available for download directly from the Bureau’s web site. No longer must an interested member of the public (press, community group, or otherwise) request your LAR from you bank directly. It’s just an easy download, in analysis-ready format, from the Bureau. This means no will no longer even know who has your information, as you did before, and makes it so much easier for your information to be used in comparison exercises by anyone with an internet connection and a spreadsheet.
The regulators have stated publicly that they will examine banks under a “good faith effort to comply” standard when it comes to the new rule, much as they did when TRID was new. But do you feel comfortable with putting data into the public domain that might not be as accurate as it could be? You might get through a HMDA data integrity exam without criticism but the fair lending impact could be a different story. How do you think responding to an inquiry about your bank’s mortgage lending patterns by saying the data is wrong will be received?
Some take the position that it won’t matter since the Bureau will soon finalize a rule to keep certain of the new data fields from the public. In its proposal, the Bureau included the applicant’s credit score, DTI, the property value, address, and AUS findings, among others, in its listing of data to exclude from public disclosure. But still leaves a large number of new data fields that will be made public, along with all-new data on open-ended lines of credit to peruse, and don’t forget that the regulators will have the full range of information at their disposal.
One thing to keep in mind is that with the dramatic change in reporting (both types of loans as well as data), it will be difficult to compare 2018 performance with 2017. It’s an apples-to-oranges exercise; trend analyses will be hindered by the changes unless banks go through the exercise of extracting the data from 2018 LARs that were also reported in previous years.
What environmental trends can we expect to see with the new data? Here are a couple thoughts:
- On average, HELOC populations tend to be wealthier and more non-Hispanic White than closed-end loan groups. Although this is a nationwide phenomenon, it would be worthwhile to determine whether this is true at your bank;
- The rise in mortgage rates and resulting end of the refinancing wave has the potential to impact the makeup of banks’ product mixes, since refis will likely make up a smaller percentage of loans (by purpose) than in previous years. This also may have an impact on race or ethnicity breakdowns or across income brackets. This is a market factor that may impact your bank’s fair lending profile rather than one made by choice. Be aware of whether and how much it impacts your bank.
- The elimination from HMDA reporting of unsecured home improvement loans will likely have a greater impact on low- to moderate-income (LMI) areas than on middle- or high-income areas. This is not a factor brought about by market changes but rather one due to changes to the regulation; nonetheless it may impact your bank’s aggregate statistics year over year.
Be aware of further regulatory changes
There won’t be any more changes since Congress and the Bureau have already spoken, you say. Except for the impending finalization of the public disclosure of data provision, this is true as far as changes to the regulation go (for now; the Bureau has said it will re-examine the new HMDA rule sometime in 2019). But keep in mind that the Bureau mandates how and what reporters submit through its Filing Instructions Guide. Although not part of the regulation per se, changes the FIG (as it’s affectionately known) are effectively changes to the rule itself. Be sure to monitor the FIG – it’s been updated four times in 2018 alone.
The 2018 HMDA rule present banks with a unique challenge, one that impacts compliance, operations, IT, investors, and anyone else involved in the mortgage loan environment. The changes promise to be far-reaching even after submission day is past. Taking the time now to put forth your best effort, including catching all the landmines, promises a greater return later.